Mesh Cell Refinement for Object Mismatch Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Comparing digital models of objects with minor differences is inefficient and resource-intensive, particularly when dealing with similar objects that share many features.
Innovation Solution
A computer system that detects and compares objects by refining mismatched cells in a mesh model based on color, texture, shape, and dimensions using a neural network, outputting location data of mismatches and generating heat maps to highlight differences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full mesh comparison is performed on objects with many common features, then measurement precision is improved, but productivity deteriorates due to excessive computational resources and time consumption
Solution Approach 1:
The patent divides the mesh comparison task into hierarchical levels: first comparing at the object level to identify potential mismatches, then refining only to cell level at mismatch locations, and finally performing detailed vertex-level comparison only where needed. This segmentation avoids unnecessary full-mesh comparisons while maintaining detection accuracy.
Solution Approach 2:
The patent applies different levels of comparison detail to different regions of the mesh. Full detailed comparison is performed only at identified mismatch locations, while matching regions receive minimal or no detailed analysis. This local quality approach concentrates computational resources where they are most needed.
2Measurement precision
If detailed cell-level refinement is performed across the entire mesh, then measurement precision is improved, but device complexity increases due to higher computational resource requirements
Solution Approach 1:
The patent performs preliminary object-level comparison before detailed cell-level refinement. This preliminary action identifies potential mismatch regions, allowing the system to avoid performing computationally intensive detailed comparisons in regions that are already known to match, thereby reducing overall computational complexity.
Solution Approach 2:
The patent extracts and focuses computational effort only on mismatched cells identified through preliminary comparison. Instead of processing the entire mesh at full detail, the system isolates and processes only the relevant portions that contain actual differences, significantly reducing device complexity requirements.
3Measurement precision
If comprehensive feature comparison (color, texture, shape, dimensions) is performed, then measurement precision is improved, but loss of time increases due to multiple analysis steps
Solution Approach 1:
The patent segments the comprehensive feature comparison into hierarchical stages: geometric structure comparison at the object level, then targeted color/texture/shape/dimension comparisons only at identified mismatch cells. This segmentation reduces total analysis time by avoiding redundant comprehensive comparisons across the entire mesh.
Solution Approach 2:
The patent performs partial comprehensive comparison rather than excessive full comparison. After identifying mismatch locations through geometric comparison, the system applies the full suite of color, texture, shape, and dimension analyses only to those specific regions, achieving comprehensive feature comparison where needed while minimizing total analysis time.
Data Source
AI summary
A system includes a computer programmed to detect a first and a second object in received image data, determine a mesh of cells on each of the first and second object surface, upon identifying a cell of the mesh on the first object mismatched to a corresponding cell on the second object to refine the mismatched cell to a plurality of cells, wherein identifying the mismatch is based on a at least one of a mismatch in a color, texture, shape, and dimensions, stop refining the cell upon determining that a refinement of the refined cell of the first object results in a refined cell that is matched to a corresponding refined cell of the second object, and output location data of mismatched cells of the first and second objects. A mismatched cell has at least one of a color mismatch, texture mismatch, and shape mismatch.